Apical four-chamber echocardiography segmentation using marker-controlled watershed segmentation

Abstract

Echocardiography provides information about size, shape, and function of heart to create the image. Apical four-chamber echocardiography can be obtained by placing the ultrasound probe at the apex of the left ventricle. Such view enables to analyze heart abnormalities. In this regard, chamber quantification is recommended to evaluate the heart volume by defining the endocardial border of the chambers. Image segmentation that subdivides an image into different regions may be suitable to analyze the chambers. In this study, we propose a Marker-controlled Watershed segmentation method. It involves Watershed Transform, internal and external markers, and morphological processing. The advantage is that it can control over-segmentation due to noise and other local irregularities. We tested the proposed method on an apical four-chamber echocardiography image and a model image. We also compared the segmentation result with an Active Contour-based segmentation method. The results show that the Watershed Transform performs better than Active Contour.